Alibaba Cloud today announced a new large language model, called Tongyi Qianwen, that it will roll out as a ChatGPT-style front end to all its business applications, including the workplace communications product DingTalk and the smart speaker and intelligent personal assistant Tmall Genie.
Currently, the software, which will support English and Chinese, is available in beta test for enterprise customers in China, Alibaba Cloud said in a press release.
Tongyi Qianwen is based on Tongyi, Alibaba’s proprietary pretrained AI framework that unifies various AI models, including models that can turn text into images and short videos.
Large language models are deep-learning algorithms or programs for natural language processing that can produce human-like responses to queries, and serve as the basis for generative AI.
Integrating AI with business applications
“Tongyi Qianwen will be integrated into all business applications across Alibaba’s ecosystem in the near future to further enhance user experience, from enterprise communication, intelligent voice assistance, e-commerce, search, to navigation and entertainment,” Alibaba said.
With the DingTalk integration, Alibaba aims to make workplace communications more efficient. Tongyi Qianwen is designed to summarise meeting notes, write email, and draft business proposals or promotion campaign plans through simple prompts, Alibaba said.
“Users can instantly create a mini application on DingTalk by photographing a draft idea written on paper,” the company said.
Tmall Genie, meanwhile, will be able to engage in more dynamic and vivid conversations with users through the Tongyi Qianwen integration. Tongyi Qianwen will let it develop and tell stories to children, provide healthy diet recipes, offer travel tips, and recommend background music for a workout, Alibaba said.
Alibaba's announcement comes on the same day that the Cyberspace Administration of China (CAC) published draft measures for generative AI services, including making providers responsible for the validity of data used to train generative AI tools.
Alibaba did not directly address the concerns mentioned in the CAC announcement, but did say that it hopes to fine-tune Tongyi Qianwen with customers’ proprietary intelligence in a secure cloud environment, so enterprises can establish tailored AI models to suit their specific business needs.
Alibaba focuses on cost-efficient AI
The company said it believes that the move will make computing more accessible and affordable for companies looking to unlock emerging opportunities in the new AI era.
“This is expected to spark a new wave of growth momentum for customers, eliminating the need for resource-intensive and expensive pre-training processes for building foundational models,” the company said.
The company emphasised that its customers and developers will have access to the model to create customised AI features in a cost-effective way.
“We are witnessing a new paradigm of AI development where cloud and AI models play an essential role. By making this paradigm more inclusive, we hope to facilitate businesses from all industries with their intelligence transformation and, ultimately, help boost their business productivity, expand their expertise and capabilities while unlocking more exciting opportunities through innovations,” Jingren Zhou, CTO of Alibaba Cloud Intelligence, said in the company announcement.
Competition from Baidu and ChatGPT
Last month, another Chinese internet services and AI giant, Baidu, announced a Chinese language ChatGPT alternative, Ernie bot. The company had said in its initial phase, 650 business partners would have access to the bot, and that it hopes to improve the bot based on feedback.
ChatGPT was launched in November and had over 1 million users within the first five days. It was launched by Microsoft-backed OpenAI. Last month, in a blog post, the company revealed the latest version of the programming language GPT-4. The GPT-4 large language model is a multimodal system that can process not just text, but images, video, or audio. The company is also looking to have the ability to reply to inputs in other languages as well.